skypilot

skypilot copied to clipboard

Change default scheduling demand of 0.5 cpu/task to 1.

1

Fixes #1045 - see that issue for the motivation of the change (UX/confusion on user's front). **This change will half the maximum number of concurrent tasks**, compared to master. This...

concretevitamin

Default scheduling demand `CPU:0.5` can be confusing

2

In `sky queue` - `[CPU:0.5]` means this task takes 0.5 cpu for **scheduling purpose**. So e.g., if the VM has 8 cores, this means we can concurrently run at most...

concretevitamin

Daniel has run into a back-compat issue, where - he had a cluster launched before our AMI upgrade; it's in stopped state - he upgraded Sky, which now includes the...

concretevitamin

sky start may unexpectedly launch a new VM

5

Mentioned by @Michaelvll in https://github.com/skypilot-org/skypilot/pull/1014#discussion_r940537044, below scenario may trigger an unexpected behavior of `sky start`. 1. User `sky launch gpunode` and got a VM in `us-west-1a` 2. User `sky stop`...

infwinston

In #978, we created programmatic APIs for basic sky commands, but the on-prem and benchmark-related functions are still in `cli.py`. We may want to move them into `core.py` as well.

Michaelvll

This PR adds document for the programmatic API, waiting for #978 TODO: - [ ] Document for the programmatic API - [ ] Example for the API

Michaelvll

Log outputs aren't all prefixed

1

Repro: ``` sky launch -c xv2test2 'git clone https://github.com/RitwikGupta/xView2-Vulcan.git && conda install --file xView2-Vulcan/spec-file.txt' ``` On 04b94b9004a, some lines are prefixed some are not: ``` mkl-service-2.4.0 | ########## | 100%...

concretevitamin

[preview] Change column limit to 100

5

Even linux is using 100-column now: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=bdc48fa11e46f867ea4d75fa59ee87a7f48be144 I feel worse and worse when looking at files like sky/backends/cloud_vm_ray_backend.py. Every two of one lines of code you need to fold it...

suquark

Fix unnecessary ssh hanging issue on Ray

5

This PR aims to fix https://github.com/sky-proj/sky/issues/849 by patching Ray autoscaler. Detailed explained in the issue. We think a simple way is to patch Ray's autoscaler by adding `self.provider.non_terminated_nodes({})` before https://github.com/ray-project/ray/blob/6d978ab10ec65da1018790f8605b5b8946e838e5/python/ray/autoscaler/_private/updater.py#L272....

infwinston

## Features - [ ] `sky admin status` - New CLI command that lists all user jobs for the administrator - [ ] Add local cluster spillover to multiple clouds...

michaelzhiluo